RAD Lab wins backing from five major IT firms

May 30, 2006

The University of California, Berkeley, today announced a new influx of industry support for its Reliable, Adaptive and Distributed systems laboratory, or RAD Lab, which will be dedicated to creating the next generation of Internet design tools. IBM Corporation, Hewlett-Packard Corporation, NTT Multimedia Communications Laboratories, Inc., Nortel Networks and Oracle Corporation have signed on as affiliate members of the Internet research lab, each pledging annual contributions of up to $170,000 for the next five years.

This corporate support reflects both the importance of the RAD Lab's work and a major shift in the way long-term university research is funded, said David Patterson, a UC Berkeley professor of electrical engineering and computer sciences and the RAD Lab's founding director. "Until recently, federal grants from the Department of Defense's Defense Advanced Research Projects Agency and the National Science Foundation covered research costs for large projects such as this," Patterson said. "But reductions in government funding of information technology research led us to seek alternative funding sources."

The RAD Lab, which launched in December 2005, is adapting advances in statistical machine learning, a form of artificial intelligence, to the task of maintaining the large distributed computing systems needed to run data-intensive Internet businesses. These distributed systems typically require hundreds of engineers for development, debugging and ongoing maintenance. The RAD Lab is designing software architecture that will automate this - eventually leading to drastic cost savings. RAD Lab technology will also make it possible for small groups or individuals to develop the kinds of Internet services that currently require large teams of engineers.

"The Internet and other distributed computing systems are critical to the functioning of day-to-day business; we rely on them to be stable 24/7," said Patterson. "Yet, today's systems are fragile. System failures and security breaches are facts of life, so the most practical solution is to focus on fast detection and recovery from failures, as well as on immediate detection and isolation or containment of security breaches. We hope to automate that with statistical machine learning."

All software and applications emerging from the RAD Lab will be made freely and openly available to the public, with source codes distributed using the Berkeley Software Distribution license. However, foundation and affiliate members of the RAD Lab will have access to the new technologies at least six months before they are made available to the general public, with research results reported first to sponsors at twice-a-year, three-day retreats. Representatives from the companies will act as consultants and provide advice for the center's participants, but will not work at the RAD Lab.

"Our member companies benefit by witnessing ideas in pre-competitive technologies at the early stages of development, and they will help point out the real world obstacles that must be overcome," said Patterson. "While the RAD Lab is committed to creating non-proprietary, open source technology, our sponsors are positioned to be early adopters, able to include RAD Lab technology in their ongoing product development as they see fit."

The RAD Lab launched late last year with $7.5 million in funding over five years from foundation members Google, Microsoft and Sun Microsystems. The research laboratory expects to receive as much as 80 percent of its support from industry. A UC Discovery Grant, a grant from the state Microelectronics Innovation and Computer Research Opportunities (MICRO) program, and another from the National Science Foundation will make up the remaining proportion of the funding for the center.

The RAD Lab was co-founded by UC Berkeley electrical engineering and computer sciences professors Michael Jordan, Randy Katz, David Patterson, Scott Shenker and Ion Stoica. Another lab co-founder, Armando Fox, just joined UC Berkeley from Stanford University. Fifteen graduate students currently participate in RAD Lab research; this number is expected to grow to 30 over the years.